Low Speed Virtual Wind Tunnel Simulation

8/10/2019 Low Speed Virtual Wind Tunnel Simulation

1/72

Low Speed Virtual Wind Tunnel Simulation For Educational Studies In

Introducing Computational Fluid Dynamics And Flow Visualization

BY

Cher-Chiang YangBS, Aerospace Engineering, University of Kansas, Lawrence 1995

Submitted to the Department of Aerospace Engineering and the Faculty of the GraduateSchool of the University of Kansas in partial fulfillment of the requirements for the

degree of Master of Science.

_______________________________Committee Chairperson, Dr. Ray Taghavi

_______________________________Committee Member, Dr. David Downing

_______________________________Committee Member, Dr. Saeed Farokhi

_______________________________Date thesis accepted


2/72

Acknowledgements

I would like to extend my most sincere gratitude to my advisor Dr. Ray Taghavi

(John E. & Winifred E. Sharp Professor of Aerospace Engineering). I would like thankhim for his constant guidance and motivation. Not only did he provide me with his

mental support, Dr. Taghavi has been a very good friend to me during my difficult times

when I went through a family crisis.

I would also like to thank Dr. Saeed Farokhi, Dr. Chuan-Tau Edward Lan, Dr.

David R. Downing. Their patience and aid have helped me tremendously in completing

this writing. The continuous encouragements from them made this completion possible.

Last but not least, I wish to thank my parents, Ee Chuang Yong and Hwee Huan

Tan. Their constant love and care for me are deeply appreciated. With the passing of my

father, I would like to dedicate this thesis to him and to all the individuals that believe in

me.


3/72

i

Abstract

Computational Fluid Dynamics tools have been around for a couple of decades

now. With the growing computing power, the speed and accuracy of these tools have

improved tremendously. The ability to visualize flow is now a common feat on the

powerful and speedy computers. Students of aerodynamics studies would benefit greatly

not only with the abilities to simulate flows, but also to visualize them. Unfortunately, to

use such tools, one has to be quite well-versed in the language of complex computational

programming.

The challenge for most aerospace or aeronautical undergraduate student is tounderstand the complicated world of aerodynamics through series of mathematical

equations. Without the ability to see how flows behave in motion, the student can only

imagine how the stall occurs over an airfoil or how the turbulent air looks like after

separation happens. In this case, a (flow separation) picture will definitely speak more

than a thousand words (or equations). Computational Fluid Dynamics offers the above

capabilities, but with a catch the user must know aerodynamics well enough so as not to

blindly believe all the computer data being spewed out is correct. The phrase garbage in,

garbage out will describe the situation most adequately if the user has little knowledge

about setting the boundary conditions or fluid properties. Also, the more complex the

simulation is, the longer it requires to compute the solution. Nowadays, as in all

processes, flow simulation is expected to work fast, if not instantaneous. However, in theworld of Computational Fluid Dynamics, typically the accuracy of the simulation is

sacrificed for the speed in obtaining the solution or vice versa.


4/72

ii

To simplify the complex mathematics involved in Computational Fluid

Dynamics, the Low Speed Virtual Wind Tunnel simulation is created. This program cuts

down on the require information from the user in order to perform a simulation. The

program is capable of taking an airfoil coordinates that is generated according to the

users specifications and provide a quick and dirty estimation of aerodynamic

characteristics like lift, drag and pitching moment. In addition to that, a pressure flow

field across the airfoil is created to show the pressure distribution of the airfoil. With

further modification to the input coordinates data, an animation of the flow is produced.

Thus this picture speaks more than a thousand words (or equations).By utilizing the speed of the computation, there are restrictions to the results

obtained. The visualizations of the flows are extremely telling but the aerodynamics

characteristics are skewed when flow separation occurs. Unsteady flow in flow separation

requires longer computing time and information to give a more complete analysis.

Therefore, results from high angles of attack in stall condition should be taken with some

skepticism.

Thus, the Low Speed Virtual Wind Tunnel simulation program remains an

acceptable tool for students who are beginners to the field of aerodynamics and

Computational Fluid Dynamics. The ability to visualize the flow field enhances the

understanding of the mathematical flow equations is undeniable. This also gives the

students an early taste of the power of Computational Fluid Dynamics in the years to

come that would play a crucial role in the ever developing aerospace industry.


5/72

iii

Table of Contents1. INTRODUCTION.................................................................................................................................... 1

2. THEORETICAL BACKGROUND AND LITERATURE REVIEW.................................................. 5

3. LOW SPEED WIND TUNNEL DESIGN ............................................................................................ 10

3.1. LOWSPEEDVIRTUALWINDTUNNELOVERVIEW.............................................................................. 12

3.1.1. 2-D Analysis with Low Speed Virtual Wind Tunnel .................................................................. 14

3.1.2. System Requirements for LSVWT.............................................................................................. 14

3.2. QUICKSTARTGUIDE TOLSVWT...................................................................................................... 15

3.2.1. Airfoil coordinates generations................................................................................................. 18

3.2.2. Panel Method Analysis.............................................................................................................. 21 3.2.3. FlowLab in 2-D usage............................................................................................................... 25

3.2.4. Starting FlowLab ...................................................................................................................... 26

3.2.5. Geometry Settings ..................................................................................................................... 27

3.2.6. Flow Conditions (Physics) Settings .......................................................................................... 27

3.2.7. Mesh Settings ............................................................................................................................ 30

3.2.8. Solve for Solution Settings ........................................................................................................ 31

3.2.9. Graphic Reports Settings .......................................................................................................... 32 3.2.10. Post-processing Analysis Settings........................................................................................... 34

3.3. ADDITIONALFEATURE INA NIMATION............................................................................................... 34

4. RESULTS AND DISCUSSIONS........................................................................................................... 36

4.1 2-D FLOWR ESULTS OF NACA 2415 .................................................................................................. 36

4.2 2-D VISUALFLOWR ESULTS OF NACA 2415 ..................................................................................... 39

5. CONCLUSION AND RECOMMENDATIONS.................................................................................. 45

5.1. LSVWT 2-D FLOW ANALYSIS........................................................................................................... 45

5.2. R ECOMMENDATIONS FOR FUTURE WORK ........................................................................................... 46

5.2.1. 2-D Analysis in LabVIEW and FlowLab................................................................................... 46


6/72

iv

5.2.2. FLUENT and GAMBIT in grid generation ............................................................................... 46

5.2.3. FlowLab in 3-D usage............................................................................................................... 47

5.2.4. OpenFlower and Gmsh ............................................................................................................. 48

6. REFERENCES....................................................................................................................................... 50

APPENDIX A: PROGRAM FLOWCHART OF LABVIEW FOR LSVWT PROGRAM LABVIEW

DETAILS OF THE PROGRAMMING....................................................................... 52


7/72

v

List of Figures

FIGURE2.1 COST AND TIME RELATIONSHIP WITH RESPECT TOCFD AND WIND TUNNELS............................. 5

FIGURE2.2 - BOEING777 DESIGN COMPONENTS AFFECTED BYCFD............................................................... 7

FIGURE2.3 NASA VIRTUALWINDTUNNEL APPLICATION........................................................................... 7 FIGURE2.4 U NSTEADY FLOW OF STREAKLINES AND TIME LINES OVER AN AIRFOIL. ..................................... 8

FIGURE3.1 LAYOUT OFKU LARGE WIND TUNNEL. .................................................................................... 10

FIGURE3.2 CURRENTKU LARGE WIND TUNNEL USER INTERFACE. ............................................................ 11

FIGURE3.3 FEATURES OF THELOWSPEEDVIRTUALWINDTUNNEL. ........................................................ 13

FIGURE3.4 STARTUP SCREEN OFLSVWT PROGRAM. ................................................................................ 15

FIGURE3.5 JAVAFOIL AIRFOIL COORDINATE GENERATION SCREEN............................................................ 16

FIGURE3.6 FLOW FIELD PLOTTING BYJAVAFOIL. ...................................................................................... 17

FIGURE3.7 FOILSIM PROGRAM IN MOTION. ................................................................................................ 18

FIGURE3.8 PANELMETHOD AIRFOIL COORDINATES GENERATION OF NACA 0012. .................................. 19

FIGURE3.9 COORDINATE GENERATION OF NACA 2415 AIRFOIL. .............................................................. 20

FIGURE3.10 AERODYNAMICS RESULTS WITH PRESSURE DISTRIBUTION ACROSS AIRFOIL........................... 21

FIGURE3.11 PANELMETHOD RESULTS OF A NACA 2415 AIRFOIL. ........................................................... 24

FIGURE3.12 FLOWLAB2-D ANALYSIS OFCLARKY AIRFOIL. .................................................................... 25

FIGURE3.13 FLOWLAB ANALYSIS MODEL SELECTION................................................................................ 26

FIGURE3.14 GEOMETRY MODULE CREATION. ............................................................................................ 27

FIGURE3.15 PHYSICS OR FLOW CONDITIONS MODULE SETTINGS................................................................ 28

FIGURE3.16 BOUNDARY CONDITION SETTINGS IN THEPHYSICS MODULE. ................................................. 29

FIGURE3.17 MATERIALS PROPERTIES IN THEPHYSICS MODULE. ............................................................... 29

FIGURE3.18 MESH SETTINGS FOR THE AIRFOIL. ......................................................................................... 30

FIGURE3.19 SOLUTION SETTINGS MODULE. ............................................................................................... 31 FIGURE3.20 GRAPHICR EPORTS MODULE. ................................................................................................. 32

FIGURE3.21 EXAMPLE OF A RESIDUALS PROGRESS WITH RESPECT TO ITERATIONS. ................................... 33

FIGURE3.22 POST MODULE TO SHOW THE RESULTS OF COMPUTATION....................................................... 34

FIGURE3.23 VONK RMN VORTEX STREET ILLUSTRATED INFLOWLAB. ................................................. 35


8/72

vi

FIGURE4.1 GENERATION OF NACA 2415 AIRFOIL COORDINATES. ............................................................ 36

FIGURE4.2 2-D RESULTS OF NACA 2415 AT ZERO ANGLE OF ATTACK . ..................................................... 37

FIGURE4.3 PRESSURE FLOW FIELD NACA 2415 AT ANGLE OF ATTACK AT10 DEGREES(TOP LEFT), 12

DEGREES(TOP RIGHT) AND14 DEGREES(ABOVE). ................................................................... 39

FIGURE4.4 PRESSURE DISTRIBUTION ON NACA 2415 AT -6 TO4 DEGREE ANGLES OF ATTACK . ................ 40

FIGURE4.5 PRESSURE DISTRIBUTION ON NACA 2415 AT 6 TO15 DEGREE ANGLES OF ATTACK ................. 41

FIGURE4.6 PRESSURE DISTRIBUTION ON NACA 2415 AT HIGH ANGLES OF ATTACK OF16 AND17 DEGREES.

............................................................................................................................................................ 42

FIGURE4.7 PRESSURE DISTRIBUTION ON NACA 2415 AT HIGH ANGLES OF ATTACK OF18 AND20 DEGREES.

............................................................................................................................................................ 43


9/72

1

1. Introduction

Computational fluid dynamics (CFD), a fast growing component in computer-

aided engineering, plays a very vital role in reducing costs and turn-around times in thedesign and development of aircraft. The CFD simulations and wind tunnel testing

represent an important phase to any aircraft design, particularly for brand new design

concepts. These complex simulations or wind tunnel results show whether the aircraft

aerodynamics behaviors are acceptable for the purpose of its design. One such example is

the Boeing 777 that utilized intricate CFD simulations extensively in its design

development of components like the wing, wing-body fairing and engine/airframe

integration. Physical testing of aircraft models in wind tunnels has remained useful in

design validation and analysis even though CFD simulations are becoming more popular

and reliable than before. To fine-tune CFD simulations and authenticate the aerodynamic

characteristics of aircraft designs, prior wind tunnel studies are compared to the simulated

CFD results of the same wind tunnel models.Since its inception in the 1950s, CFD has matured progressively and advanced

greatly especially in the past two decades. In the early days of CFD, supercomputers were

needed to process the long and tedious CFD calculations. Later, researchers moved away

from the expensive and limited availability of supercomputers by running workstations in

parallel processing to accomplish the CFD tasks. With the emergence of powerful

personal computers in the last two decades, most of the CFD simulations can now be

achieved relatively quickly compared to the workstations.

Two features of the CFD outshine wind tunnel testing and the element of cost is

one of such advantages. During the preliminary aircraft design phase, wind tunnel models


10/72

2

undergo multiple modifications. These modifications, which can lead to higher costs, are

necessary in order to optimize design configuration or allow iteration changes.

Fortunately, CFD simulations do not require these costly and time-consuming model

modifications. There is no expensive model alteration to carry out or down time in the

wind tunnel while the model is being fixed. These CFD simulations can apply changes to

the virtual models as quickly as they can be modified in the computers to obtain new

results. This time saving benefit is another edge that CFD simulation has over the

traditional wind tunnel testing. In the same amount of time needed to conduct a wind

tunnel testing, many simulations could be completed to produce far more extensiveresults and detailed flow field information that wind tunnel results are incapable of

showing. For full configuration aircraft models, these extensive results can show the

detail flow field interaction of the wing-fuselage interface whereas the wind tunnel results

can only present the overall aerodynamics behaviors. In the design phase, especially in

the preliminary stage, it would be impractical to study several major configuration

changes without the use of CFD. The ability to obtain results with CFD in a short amount

of time stands out against wind tunnel testing that requires time to create or modify a

model.

Although CFD offers quicker solutions, this does not mean that wind tunnel

testing is obsolete. Wind tunnel data still play a key role in the design validation of

configurations, but CFD simulations can take a step further with complex configurations

analysis and enhance rapid prototyping capability. There is still a considerably strong

need for basic wind tunnel experiments to validate CFD data in areas such as flow

stability, 3-D boundary layers and flow separation characteristics. Through data


11/72

3

validation, CFD simulations accuracy will steadily improve and then will be capable of

simulating results even for a conceptual design before going into a wind tunnel. When

used in conjunction with wind tunnel testing, CFD can help to determine and refine wind

tunnel experimental data due to interferences from the tunnel walls and model mounting

system. Thus, this creates a synergistic use of CFD and wind tunnels that will aid the

development of more effective and reliable simulations.

As powerful as CFD simulation can be, it also have weaknesses and pitfalls if the

user applies it inappropriately. The simulations sophistication is the strength and the

weakness that presents to researchers or aerodynamics students with little or no CFDexperience will face. Most current CFD tools are too difficult for new users with limited

aerodynamics knowledge to perform simulations by themselves. These non-CFD users

will also be looking at a rather challenging task in grid generation. With knowledge in

fluid dynamics but not in computational mechanics, they do not know which grid to use

or how to specify minimum spacing.

To address such difficulties and make CFD a relatively user-friendly tool, the

current project attempts to combine the advantages of CFD and wind tunnel testing to

provide a unique educational and experimental platform for aerodynamics study. The

Low-speed Virtual Wind Tunnel (LSVWT) computer program will simulate the KU large

wind tunnel that is capable of running at a maximum velocity of 185 miles per hour and

allow users to have a hands-on experience of a typical wind tunnel operation to obtain

aerodynamics results for preliminary design analysis.

The goal of this Low-speed Virtual Wind Tunnel (LSVWT) program is to apply

the simplified Computational Fluid Dynamics (CFD) calculations while providing an


12/72

4

easy and intuitive interface with a wind tunnel. Prior research in the past has focused on

either wind tunnel testing or CFD separately. This simulation will combine the ability of

swift design changes in CFD with the reliability and the repeatable, trusted wind tunnel

results.

Continuing from this introductory chapter, Chapter 2 provides a literature review

based on the wind tunnel simulation and its capabilities. In Chapter 3, the detailed setup

for the CFD will be described, along with the grid generation for a virtual airfoil model.

A comparison and validation of the 2-D airfoil CFD simulated results is discussed in

Chapter 4. In the same chapter, a complete visualization of the CFD results at differentangles of attack is also presented. Finally, Chapter 5 contains conclusions and

recommendations for further enhancements on the LSVWT.


13/72

5

2. Theoretical Background and Literature Review

Throughout the design phase of a vehicle, improvements and modifications

usually occur that lead to design changes. Although design iteration is not completelynew to CFD, its use in conjunction with wind tunnel data is not often adequately explored

enough. Previous studies of CFD did not include wind tunnel data, and focused on

reducing the time needed to solve for accurate solution convergence. The numerical wind

tunnel researched by Bell1 demonstrated the computational mechanics knowledge that is

required to carry out a CFD simulation of a wind tunnel and the focus was to speed up the

time to obtain the solution. It is important to solve for the solution convergence in a

relatively short time so that the vehicle design is examined and improved in the same

amount of time it takes to conduct a single wind tunnel experiment. This is illustrated in

Figure 2.1.

Figure 2.1 Cost and time relationship with respect to CFD and wind tunnels.2


14/72

6

Saving time is one of the key features in using CFD. However, no matter how fast

a CFD solution is presented, the results do not bear much technical value if the CFD

modeling is not supported by any wind-tunnel-based experimental data. This type of

numerical simulation would only provide the insight for better mathematical code

optimization, and not improving the vehicle design significantly.

To create synergism between CFD and wind tunnel testing, several accurate tests

are used to form the building blocks of the validation of the computational results. The

validation of data starts with a 2-D aerodynamics analysis on an airfoil, with the focus on

the aerodynamics characteristics and the stall behavior. Once validated, an airfoil modeldesign can be inserted into the CFD simulation to be tested for aerodynamic behavior and

flight characteristics. Design corrections can be made so that the desirable behavior and

characteristics are obtained in the simulation. This process of design optimization creates

a refined wind tunnel model vehicle that should perform very well in aerodynamic terms

and demonstrate desirable flight characteristics in the testing. Tinoco2 showed evidences

of the conjunction usage of CFD and wind tunnel testing in influencing and optimizing

Boeing 737, 757, 767 and 777 component designs. The components that were influenced

by CFD are shown in Figure 2.2. This synergistic use, however, needed experienced CFD

users.


15/72

7

Figure 2.2 - Boeing 777 design components affected by CFD.2

For non-CFD users, Fujii and Miyaji3 of Japan created a web-based CFD tool to

process grid generation, flow simulation and visualization with limited body

configurations. But their application was narrowed to rocket and rocket nozzle

configurations. There is a handful of commercial CFD software available but are too

complicated for inexperienced users.

Figure 2.3 NASA Virtual Wind Tunnel application.6


16/72

8

NASAs Glenn Research Center has been developing the application of virtual

tunnels for many years and even more so recently with the improvement of computing

power. One of these virtual tunnels named Immersive Connection to RemoteWindTunnel is shown in Figure 2.3. However, due to the complexity of the programs, other

than the CFD specialists, most of the communities do not have easy access to these CFD

tools. For instance, the state-of-the-art Unsteady Flow Analysis Toolkit (UFAT)

developed at NASA Ames Research Center is a pioneering tool in visualizing unsteady

flow simulations.

The UFAT program can plot streaklines and time lines that are time-dependent

particle tracing techniques. Those techniques are very effective for visualizing unsteady

flows like an unsteady flow data surrounding an oscillating airfoil shown in Figure 2.4.

However, to obtain those results, users are required to understand and setup the complex

conditions in CFD. As powerful as UFAT is, it is not a program that is easily understood

by any user.

Figure 2.4 Unsteady flow of streaklines and time lines over an airfoil.7


17/72

9

A simple aircraft-related CFD simulation is needed and thus the Low-speed

Virtual Wind Tunnel (LSVWT) concept is born.

This project will show how the LSVWT handles the two main pieces of the

program in grid generation and flow simulation. The panel method approximation is

applied in the 2-D airfoil analysis with the capabilities of generating coordinates for 4 and

5-digit NACA airfoils. FlowLab, a commercial software, enhances the flow field analysis

and a second source to verify the panel method result. FlowLab is the simplified version

of its more complex parent FLUENT. The 2-D analysis in FlowLab uses the same

FLUENT commercial code that utilizes the Navier-Stokes equations to solve for varioustypes of flow and turbulence models. GAMBIT, a grid generation program, works with

FLUENT to discretize the domain to form structured and/or unstructured grid in order to

solve the Navier-Stokes equations for inviscid or viscid flow.

The goal in creating LSVWT is to provide non-CFD researchers or students a user

interface that quickly set-up to CFD analysis. With the LSVWT controls modeled after

the KU large wind tunnel user interface, students can familiarize themselves with the

actual large wind tunnel through the usage of LSVWT.


18/72

10

3. Low Speed Wind Tunnel Design

The Low-speed Virtual Wind Tunnel (LSVWT) program consists of a 2-D

analysis portion. LabVIEW is the primary software used in designing the user interface to perform simulated wind tunnel testing and 2-D flow analysis. LabVIEW is chosen to

build the program because the KU large wind tunnel uses the very same software for its

wind tunnel data acquisition. The commonality in the wind tunnel controls is reflected in

LSVWTs control panel layout. Figure 3.1 shows the schematics of the large wind tunnel

diagram.

Figure 3.1 Layout of KU large wind tunnel.8

The subsonic large wind tunnel is closed circuit and has a 36" by 51" test section

and a maximum speed of 185 mph. It is equipped with a six-component strain-gage

balance and a PC-based LabVIEW data acquisition system. The user can set the test

section velocity with a remote throttle control. The LabVIEW user interface provides the

real time monitor that shows the aerodynamics characteristics and coefficients as shown

in Figure 3.2.


19/72

11

Figure 3.2 Current KU large wind tunnel user interface.

The interface also displays the tunnel velocity (in feet per second) andtemperature (in Fahrenheit). The inputs required by the user are:

Atmospheric pressure in inches of mercury.

Mean aerodynamic chord of model in inches.

Wing span of model in inches.

Reference area of model in inches squared.

Each desired angle of attack in degree.


20/72

12

The following are its characteristics:

Table I Characteristics of the KU low-speed large wind tunnel.

Characteristics Data

Tunnel type Closed circuit, single return

Test section type Closed, rectangular shape

Test section size W = 51, H = 36, L = 70

Power source 300 hp constant rpm electric motor

Fan type Four-bladed, variable pitch fan

Maximum test section velocity 185 mph

Turbulence factor 1.1

Contraction ratio 0

Test section sidewash* Maximum +1.8, average +1.3

Test section downwash* Maximum 1.3, average 0.4

Test section pressure Atmospheric

* The average values calculated in side wash and downwash are for the cross section

3.1. Low Speed Virtual Wind Tunnel Overview

The LSVWT can be used as an educational tool to introduce aerodynamics study

to students who are new to aerodynamics and CFD. With the control panel layout of the

LSVWT being almost identical to that of the KU large wind tunnel, the students will be

familiar with the large wind tunnel controls before they get to use it aerodynamics

characteristics analysis. The 2-dimensional analysis provides lift, drag and pitching

moment characteristics in relation to change in angle of attack and airspeed. The students

can correlate the lift, drag and pitching moment equations to varying angle of attack and


21/72

13

airspeed when they can see the immediate changes in those values as they change the

testing conditions. On top of the aerodynamics parameter values, a pressure distribution

across the 2-dimensional airfoil will be shown in real time. The students will have a

better understanding of pressure changes on the airfoil surface with the different angle of

attack settings.

For a more advanced study on 2-dimensional aerodynamics, FlowLab is called up

to provide a contour plot of the flow field around the tested airfoil. Velocity vectors and

entire flow field pressures are shown in full color to represent the wide range of values

from freestream to surface of the airfoil locations.

Figure 3.3 Features of the Low Speed Virtual Wind Tunnel.

Figure 3.3 shows a general roadmap of the LSVWT features where 4 different

modules work together to provide an uncomplicated and general CFD tool.

Low SpeedVirtual Wind

Tunnel LabVIEWPanel Method

FlowLab

JavaFoil NACA type airfoil results for C l,C d, C m in static flow field

FoilSim

Cylinder results for C l, C d, C mplus flow field analysis

Non-standard type airfoil resultsfor C l, C d, C m in dynamic flowfield

Any type airfoil results for C l, C d,Cm in dynamic flow field

Airfoil results for C l, C d, C m plusflow field analysis


22/72

14

3.1.1. 2-D Analysis with Low Speed Virtual Wind Tunnel

The LSVWT program has the following main features:

The control panel layout resembles the actual layout of the KU large wind tunnel

data acquisition system.

In 2-D airfoil analysis, the lift, drag, and pitching moment along with surface

pressure distribution are presented.

In the flow analysis, the contour, vector velocity and streamline can be shown.

All the CFD work, including grid generation and simulation, can be done with

limited knowledge of CFD.

The flow regime is restricted to subsonic range and templates of several model

setups are prepared in FLUENT so that they can be studied in FlowLab.

The templates are modular in design and more can be created (using FLUENT

and GAMBIT by experienced users) to expand the library of CFD analysis.

3.1.2. System Requirements for LSVWT

The computer system requirements for LSVWT are dictated by the sum of

programs involved in CFD analysis. The main requirements are summarized as follows:

A Pentium 4 or equivalent processor is recommended. A minimum hard disk

space of 800 MB running on Windows 2000/XP or later is needed. The computer

must be able to use Internet Explorer 5.5 with Service Pack 2 or later.

A computer with LabVIEW version 7.1 installed.

A computer with MATLAB version 5 installed


23/72

15

A computer that is able to launch internet browsers like Internet Explorer or

Mozilla Firefox to use Java applets.

Networked computer that has FLUENT, FlowLab and GAMBIT installed and the

license to execute all three programs in Exceed X-Server environment.

3.2. Quick Start Guide to LSVWT

With the system requirements mentioned earlier satisfied, the LSVWT program

can be started. Click on the LSVWT icon will launch the program and the user will see

the menu screen as shown in the following figure.

Figure 3.4 Startup screen of LSVWT program.


24/72

16

The user has a choice of 2 simple web-based flow analysis modules and 2 more

advanced simulations. An airfoil grid generation program is included to aid the more

advanced 2-D flow analysis programs. The JavaFoil8 module is developed by Martin

Hepperle that utilizes potential flow and boundary layer analysis without taking flow

separation into account. FoilSim9, developed by a team led by NASA scientist Tom

Benson, provides simple flow visualization of airflow over an object like an airfoil

without the complexity of flow separation. The Airfoil Generation and 2-D Flow modules

capabilities will be discussed with further details in the chapters to come.

Clicking on the JavaFoil button will launch the module in a separate window.Javafoil is equipped with its own airfoil generation so that it can proceed with the flow

over airfoil analysis. In Figure 3.5, a NACA 2415 is generated with the coordinates

shown.

Figure 3.5 JavaFoil airfoil coordinate generation screen.


25/72

17

This interactive web-based program is capable of calculating the velocity and

pressure distribution across the chosen airfoil. As soon as an airfoil shape is generated

from its airfoil library, the program can generate velocity and pressure distribution with

the default setting for angles of attack. A table of results that includes the lift, drag and

pressure coefficients is calculated and can be printed. One of the useful features of

JavaFoil is the plotting of the flow field as shown below in Figure 3.6. JavaFoil is simple

to use but it does not handle flow separation issues.

Figure 3.6 Flow field plotting by JavaFoil.

To view a moving flow field, another web-based program FoilSim will fill the

need. FoilSim utilizes animation to further enhance the flow visualization. Clicking onthe FoilSim button from the LSVWT main menu will launch a separate window to the

web-based program as seen in Figure 3.7. This program functions very much like


26/72

18

JavaFoil but it runs in constant simulation. FoilSim runs more like a demonstration

because the airfoil shape model does not follow any NACA specification in details.

Figure 3.7 FoilSim program in motion.

However, with the ability to change the angle of attack, camber and thickness of

airfoil on the fly, the results are shown instantaneously with the lift curve slope, pressureand/or velocity distribution. The flow visualization is quick but not extremely accurate

without taking flow separation into account.

3.2.1. Airfoil coordinates generations

To initiate the 2-dimensional analysis, the coordinates of an airfoil must be

defined. The airfoil coordinates for NACA 4, 5 and 6-digit airfoils are generated within a

LabVIEWs subprogram using the NACA equations involving polynomials that relate to

airfoil camber line and thickness distribution. The NACA 2415 and 0012 airfoils are

chosen as the subject of case study because these shapes are used in the Cessna 210 wing


27/72

19

and horizontal stabilizer. The published 2-D aerodynamics results of these airfoils are

also readily available in Pope4.

Figure 3.8 Panel Method airfoil coordinates generation of NACA 0012.

The coordinates generated come in two columns namely in the Cartesian X and

Y arrangement. To simplify the program setup, the number of panels to represent an

airfoil is set to 50 for this analysis. There will be 25 upper surface X-Y pair coordinates

and another 25 for the lower surface as shown in Figure 3.8. When the user clicks on the

Generate Airfoil button that is highlighted in Figure 3.8, a new window will pop up with

the interface to generate a NACA airfoil shape. If the airfoil is symmetrical, there is a

button (next to the Generate Airfoil) for that option. This helps to cut down on the points

generated by simply mirroring the upper half of the airfoil shape.


28/72

20

Figure 3.9 Coordinate generation of NACA 2415 airfoil.

The user can select the type of NACA airfoil from the pull-down menu from the

top left corner. Once selected, the corresponding section number for the desired airfoil

shape should be entered. The number of points required for this setup is picked to be 50.

More points will provide more details to the airfoil shape but at the same time, the

computation time will increase as well. 50 points are chosen so as to strike the balance

between airfoil shape details and speed of computation in the aerodynamics calculations.

Once the airfoil coordinates are ready, the user can click on Get data button to get the 2-D

results as shown in Figure 3.10.

Type of Airfoil

Section Number

Number ofPoints


29/72

21

Figure 3.10 Aerodynamics results with pressure distribution across airfoil.

The coordinates are fed into the Panel Method solver to obtain aerodynamics

results. The values of lift, drag and pitching moment coefficients are calculated and

presented in numeric and a pressure distribution graph. The pressure distribution

locations on the airfoil are the same locations of the coordinates generated earlier in

Figure 3.8. By changing the Angle of Attack knob and clicking on the Get data button,

real-time results of the coefficients and the pressure distribution will be updated.

3.2.2. Panel Method Analysis

The Panel Method calculates the velocity distribution along the surface of a

defined panel from an airfoil. The Kutta condition is applied here for linear and steady

Angle of Attackknob


30/72

22

flow for the boundary condition setup. The governing equation (Laplaces equation or the

linearized form in compressible flow) is recast into an integral equation. This integral

equation involves quantities such as velocity, only on the surface, whereas the original

equation involved the velocity potential all over the flow field. The surface is divided into

panels or boundary elements, and the integral is approximated by an algebraic

expression on each of these panels. A system of linear algebraic equations result for the

unknowns at the solid surface, which may be solved using techniques such as Gaussian

elimination to determine the unknowns at the body surface.

The equations governing 2-D, incompressible, irrotational flow are:

Continuity: 0=+ yv

xu

(1)

and, irrotationality:

u y

v x

= 0 (2)

One can define a velocity potential such that

x u y v

= =; (3)

This equation satisfies the irrotationality. Continuity equation becomes:

2

2

2

2 0 x y+ = (4)

One can also define a stream function such that

v xu y==

- ; (5)

which yields the following relation:

2

2

2

2 0 x y+ = (6)


31/72

23

Equations (3) and (6) are each called Laplaces equation.

In subsonic compressible flow, the potential flow equation is modified to give the

following, approximate equation:

( )1 022

2

2

2 + = M x y

Assuming one can solve for either the velocity potential or the stream function

and its derivatives (which yield the flow velocities u and v), the pressure can be

computed for incompressible flows, from Bernoullis equation as:

( ) p u v p V + + = +

12

12

2 2 2

From pressure, a non-dimensional quantity called the pressure coefficient may be

computed:

C p p

V

u vV

V V p

=

= +

=

12

1 12

2 2

2

2

2

where u and v are Cartesian components of velocity V.

With angle of attack chosen by the user, the lift, drag and pitching moment

coefficient are calculated. The surface pressure distribution is also computed and then

presented visually by LabVIEW. The result of such typical calculation is shown in Figure

3.11.


32/72

24

Figure 3.11 Panel Method results of a NACA 2415 airfoil.

Through the change of angle of attack, a 2-D airfoil lift curve slope and drag polar

can be plotted. The pressure distribution can also be shown in the corresponding locations

of the coordinates generated for the airfoil. However, the results do not take into account

of flow separation at high angle of attack (more than 12 degrees) settings. Even though

the stall characteristics do not match up favorably with Popes findings at the high angles

of attack, the results do show that the aerodynamics coefficients are similar at the lowerrange of angles of attack.


33/72

25

3.2.3. FlowLab in 2-D usage

Figure 3.12 FlowLab 2-D analysis of Clark Y airfoil.In the same way the Panel Method is activated, FlowLab can be launched from

the LSVWT program. Depending solely on the template used, FlowLab performs quick

and limited CFD analysis of the 2-D subject. The subject can range from an airfoil like

the NACA 4- or 5-digit series to a simple cylinder. Figure 3.12 shows an example with

Clark Y airfoil. The user may adjust the characteristic length of the test subject, Reynolds

Number, types of mesh and grid. These choice selections are completely fixed during the

creation of the template. Once the solution converges, pressure and velocity profile of the

entire flow field can be plotted and saved for further analysis or comparison.


34/72

26

3.2.4. Starting FlowLab

Click on the Start FlowLab button will launch the program and a screen following

that, shown below, provide models to be analyzed. The models used in the LabVIEW

module are readily converted from LSVWT when needed.

Figure 3.13 FlowLab analysis model selection.

Once that is chosen, FlowLab will load the model and proceed to the main

program (shown in Figure 3.14) that will consists of the flow field/ grid screen, controls

for analysis and overview notes on the operation of the case study.


35/72

27

3.2.5. Geometry Settings

Figure 3.14 Geometry module creation.

From the Geometry module, the chord length can be determined. The length

limitation is set accordingly to the grid that is generated around it. The user can choose to

enter the length in metric or British units. Once the chord length is set, the user can click

on Next to proceed to the next module that sets the physics of the flow.

3.2.6. Flow Conditions (Physics) Settings

Flow conditions are chosen in this module as shown in the Figure 3.15. The

condition can be set to either inviscid or viscous. For the simplicity of calculations,


36/72

28

inviscid condition is usually selected. Boundary condition and materials can be selected

to reflect the fluid of the flow.

Figure 3.15 Physics or flow conditions module settings.

For the boundary conditions as seen in the Figure 3.16, the far field pressure and

temperature can be set for the flow. This is also where the user can set the velocity (in

terms of Mach number) of the flow and the angle of attack for the airfoil. The wall

roughness is typically ignored for the quick calculations.


37/72

29

Figure 3.16 Boundary condition settings in the Physics module.

After the boundary condition is set, the density and viscosity of the flow can bechosen by activating the Materials option. In that option, the user can also determine the

thermal conductivity, specific heat and molecular weight of the fluid.

Figure 3.17 Materials properties in the Physics module.

Once the Physics setup is complete, the Reynolds Number can be calculated when

the Compute button is pressed. Then, the user can move on to the next module by

clicking on the Next button.


38/72

30

3.2.7. Mesh Settings

The intensity of meshes is chosen here will affect the solving time directly. The

user has the straightforward options under Mesh Density to use Fine, Medium or Coarse

setting. The Cell Count is for the user to gauge the complexity of the proposed mesh. The

higher the cell count is, the longer it will take for the program to complete the

calculations. The wall function is typically set to standard unless specified otherwise.

Figure 3.18 Mesh settings for the airfoil.

Once the mesh is selected, the user must click on Create to generate the newly

selected type of mesh to be used. A progress bar will appear at the top and once that

disappears, the user can proceed to the next module by clicking on the Next button.


39/72

31

3.2.8. Solve for Solution Settings

The number of iterations is set here that will determine the time to take to solve

the computations. From the screenshot below, the user will enter the number of iterations

desired and also the convergence limit of the solution. The driving factor of time required

will be the convergence limit set because the program will only stop computing once the

limit is set unless it has reached the iterations count. If the iterations are set too low, the

solution may not converge and thus yielding no results from the calculations. For a

decent convergence, the limit is set to 0.0001.

Figure 3.19 Solution settings module.


40/72

32

The user will need to click on Iterate to start the program to solve for the flow

field solution. After the iterations are completed, the graphic reports will be available in

the next module.

3.2.9. Graphic Reports Settings

Once iterations are completed, the graphic reports are available. In this module,

the aerodynamics coefficients are shown as seen in the figure below. The user will also

have the options to plot graphs of the residuals calculation progress (see Figure 3.21) and

the pressure distribution of the airfoil for that particular velocity and angle of attack.

Figure 3.20 Graphic Reports module.


41/72

33

Figure 3.21 Example of a residuals progress with respect to iterations.


42/72

34

3.2.10. Post-processing Analysis Settings

Figure 3.22 Post module to show the results of computation.

This is where the flow field can be plotted and shown depending on the users

choices. As shown in the Figure 3.22, the flow field can be plotted using (pressure or

velocity) contour, vector or streamlines.

3.3. Additional Feature in Animation

When the CFD model in FlowLab is configured in another grid format, time steps

of the flow can actually be seen. The pictures of the flow are captured in the time steps

specified by the user. An example of such application is shown in Figure 3.23 where a


43/72

35

sphere is subjected to a relatively low Reynolds Number flow to induce the Von Krmn

vortex street.

Figure 3.23 Von Krmn vortex street illustrated in FlowLab.

This advantage of such feature is that user no longer has to imagine the flow

behavior of unsteady flow. The unsteady flow is captured vividly in FlowLab where it

can be exported as a series of pictures or as an animation.


44/72

36

4. Results and Discussions

4.1 2-D Flow Results of NACA 2415

With the known mathematical equation to create the NACA 2415 airfoil, the

coordinates are generated as shown in Figure 4.1.

Figure 4.1 Generation of NACA 2415 airfoil coordinates.

The results of lift, drag and pitching moment coefficients are shown as seen in

Figure 3.9. Each of the run here gives the results for one angle of attack setting. To

produce a lift curve slope and the variation of drag and pitching moment due to different

angle of attacks, multiple runs are needed.


45/72

37

Figure 4.2 2-D results of NACA 2415 at zero angle of attack.

With each run of different angle of attack, a set of values for lift, drag and

pitching moment coefficients are calculated, along with a pressure distribution chart. An

angle of attack range from -6 to 18 is selected, with an interval of 2 degrees between each

angle. The computed results and the NACA findings are compiled as follows:

-0.6

-0.5

-0.4

-0.3-0.2

-0.1

0.0

0.1

0.2

0.30.4

0.5

0.6

0.7

0.8

0.9

1.01.1

1.2

1.3

1.4

-8 -6 -4 -2 0 2 4 6 8 10 12 14 16 18 20 22

Ang le of Attack , (deg.)

L i f t C o e

f f i c i e n

t , C

L

( ~ )

CFD

NACA data

Figure 4.3 Lift curve slope of NACA 2415.


46/72

38

From Figure 4.3, the lift curve slope obtained from the computation is fairly close

to the NACA results. At the lower angle of attacks from -6 to 2 degrees, both sets of data

are almost on top of each other. They show about the same values for lift coefficient forthose angles of attack. However, the discrepancy starts right after 2 degrees. From 2 to 10

degrees, the difference between the computed and NACA values begins to grow. The

computed values show a steeper linear increase compared to the NACAs. At this point,

the computed result also show that the airfoil has reached its stall around 10 degrees for

a lift coefficient of 1.22. As for the NACA data, the airfoil stalls at a later angle of attack

of 12 degrees and it has a slightly lower lift coefficient of 1.2.

The lift curve slopes also show that the stall characteristics are different between

the computed and experimental data from NACA. The computed lift curve slope has a

more gentle stall behavior as the lift coefficient tapers out from 1.22 to 1.15 at around 16

degree angle of attack. In contrast, the NACA data that stalls later at 12 degrees angle of

attack drops its lift coefficient from 1.2 to 0.85 when it reaches 16 degrees. This show the

computational results will be fairly accurate up to the point where flow separation may

have occurred.

A closer observation to this stall difference can be made using FlowLab. By

calling up FlowLab, the flow field at angles of attack from 10, 12 and 14 are shown in

Figures 4.3. Note that the airfoil is referenced always horizontally, which is the default

orientation of FlowLab.


47/72

39

Figure 4.3 Pressure flow field NACA 2415 at angle of attack at 10 degrees (top left),12 degrees (top right) and 14 degrees (above).

From Figure 4.3, the unsteady flow (blue color region) is seen growing from each

angle of attack progression. The flow separation at the trailing edge of the airfoil is barely

noticeable at 10 degrees although this is the stall angle of attack for the computational

result. Looking at the 12 degrees angle of attack, the separation has crept forward to morethan half of the airfoils chord length. Thus, this corresponds to the drop in lift coefficient

value from 1.22 to 1.19. As the angle of attack increases further to 14 degrees, the

unsteady flow has almost reached the leading edge of the airfoil and therefore produces a

lower lift coefficient value of 1.10.

4.2 2-D Visual Flow Results of NACA 2415

Closer observations of the results of the NACA 2415 can be seen here. The visual

results will be shown in the progressing order of the angle of attack changes starting from

-6 to 14 degrees with a 2-degree step. From 15 to 20 degrees angle of attack, the


48/72

40

increment will be 1 degree to demonstrate drastic changes in pressure at high angle of

attack.

6 degree angle of attack 4 degree angle of attack



Figure 4.4 Pressure distribution on NACA 2415 at -6 to 4 degree angles of attack.


49/72

41




Figure 4.5 Pressure distribution on NACA 2415 at 6 to 15 degree angles of attack.


50/72


51/72

43

18 degree angle of attack

20 degree angle of attack

Figure 4.7 Pressure distribution on NACA 2415 at high angles of attack of 18 and 20degrees.


52/72

44

The visual results of the NACA 2415 shows a steady low pressure (indicated with

blue color) buildup on top of the airfoil as the angle of attack increases. At the 16 degree

angle of attack, flow separation occurs and thus creating unsteady airflow. Once theairfoil experiences unsteady flow, the computation on the aerodynamics characteristics

like lift, drag and pitching moment will start to increase at a higher rate. At this point,

only the visual results are considered trustworthy. From Figure 4.6 and 4.7, the flow

separation begins to propagate downstream and create larger wake as the airfoil increase

in angle of attack. The turbulence due to the wake generates higher aerodynamics

characteristics, in particular lift coefficient. Thus, the stall characteristics of the airfoil

would be inaccurate but the visual results do provide an insight to when and how the

turbulent flow occurs and behaves.


53/72

45

5. Conclusion and Recommendations

5.1. LSVWT 2-D flow analysis

Both web-based modules of JavaFoil and FoilSim offer quick visualization ofairfoil and flow analysis. However, each has its shortcomings; JavaFoil can only perform

analysis with each time a button is pushed and FoilSim is not able to generate specific

any airfoil shapes including NACAs. Also, both JavaFoil and FoilSim are not suitable

for flow that has separated.

Fortunately, the other 2-D modules are able to compensate on those flaws. The 2-

D Flow Over Airfoil module can provide aerodynamic characteristics continuously once

it has started. The 2-D Flow Visualization module through FlowLab is able to show the

flow separation. However, the values of the lift, drag and pitching moment coefficients

are slightly different as seen in the case study of NACA 2415 airfoil. At the 10 degree

angle of attack, flow separation may have occur which indicates the decrease in lift

coefficient as shown in Figure 4.3. The CFD stall behavior is more gentle compared tothe NACA data. This shows that FlowLab may have predicted the stall angle correctly,

but the post-stall behavior does not match up with NACAs results.

Overall, the results are satisfactory for non-CFD researchers or students learning

the basics of aerodynamics and CFD. The flow visualization will help users to understand

the flow behaviors, especially when the flow is unsteady.


54/72

46

5.2. Recommendations for future work

To further enhance the capabilities of LSVWT, an extensive 3-D model module

should be developed. There are very few CFD tools that are user friendly for CFD beginners and the time required to solve the CFD calculations are long (as in many hours,

if not days). The following are the possible future development to consider.

5.2.1. 2-D Analysis in LabVIEW and FlowLab

Although there is an Airfoil Generation option, the generated airfoil can only be

used in LabVIEWs Panel Method module currently. To use the same airfoil for FlowLabanalysis, the user need to edit the airfoil coordinates data text file manually so that

FlowLab can use it. A simple computer programming in LabVIEW could automate this

process so that the user can have the airfoil coordinates data converted for FlowLab use

with a click of a button.

With a better understanding of the condition settings and tweaking in FlowLab,

the post-stall behavior could be modeled more closely to the experimental NACA results.

With these tweaks, the CFD simulation will be more reliable and able to provide even

more convincing aerodynamic characteristics values.

5.2.2. FLUENT and GAMBIT in grid generation

GAMBIT is the primary software in creating geometry and mesh generation for

FLUENT. Virtual models can either be built in GAMBIT or imported from other

computer-aided design (CAD) programs such as CATIA or Pro/E. Other than the

common Cartesian type mesh, GAMBIT is also capable of producing triangular surface

meshes and tetrahedral volume meshes.


55/72

47

Once the geometric shape and mesh is completed, FLUENT performs the CFD

analysis. Conditions of the test environment and model are tested and this is where the

template is created with the choice selections that are later available to FlowLab. The

template creation should be left to the experienced CFD users and the non-CFD users can

benefit from the simplified case studies of the CFD templates to be used in FlowLab.

To get a 3-dimensional analysis, FLUENT can be used to provide the detail flow

field representation. However, FLUENT does require more than minimal proficiency in

CFD knowledge in setting up for the computation. A library of models will be provided

so that users do not need to know much in model grid generation using GAMBIT.The Cessna 210 that has been tested annually in the KU large wind tunnel would

be a good candidate as the test subject in the 3-dimensional CFD analysis. The results can

then be combined with the wind tunnel experimental data to provide a more complete

picture of the models flow behavior in the wind tunnel. The CFD results in the flow

analysis, coupled with velocity and pressure will compliment the experimental data in

order to provide details on the relationship between aerodynamics and flow

characteristics. This should verify whether the CFD setting is correct.

5.2.3. FlowLab in 3-D usage

FlowLab, once again, can play the pre- and post-processing role of CFD analysis

as in the 3-D scenario. The results obtained in the 3-D setup will be similar to the 2-Dsituation with the calculated velocity contour, vectors and streamline except for that

difference in which the flow field is in 3-D. The user can take a zoom-in close look at the


56/72

48

flow interference with the virtual test model. FlowLab can provide a simply user interface

to analysis the flow compared to FLUENT that is much harder to use.

5.2.4. OpenFlower and Gmsh

OpenFlower (Open Source Flow solver) was a joint effort and product of some

CFD research engineers launched in 2004. The open-source nature of the software

provides a public platform in which all levels of CFD users can contribute to improve it

so that the increasing CFD industrial need can be met. This publicly free software is

mainly devoted to the resolution of the turbulent unsteady incompressible Navier-Stokesequations. The grid generation portion is managed by another open-source software

called Gmsh that creates 3-D finite element mesh and works with OpenFlower in pre- and

post-processing of solutions.

OpenFlower is a free open-source finite volume CFD software, mainly devoted

to the resolution of the turbulent incompressible Navier-Stokes equations, with scalar

transport.5 It is a command line solver that would handle geometry and mesh generated

by Gmsh, a mesh and grid generator. The current stage of development in 3-D flow

analysis is not yet proven to be stable in the aircraft application. Thus, the results from

OpenFlower are strictly for evaluation purposes, not for close comparison.

Gmsh, the pre- and post- processor, is consisted of four modules: geometry, mesh,

solver and post-processing. It is an automatic 3-D finite element grid generator (primarilyDelaunay) with a build-in CAD engine. Since OpenFlower is a command line program

(meaning there is not any user interface), the results are shown using Gmsh. One


57/72

49

significant limitation of Gmsh is that it can only post-process 3-D results from

OpenFlower.

If the more expensive FLUENT software is unavailable, the alternate choice

would be OpenFlower and Gmsh. Although these two programs are free to download,

they are not thoroughly tested like FLUENT. The up side to the free programs is that

there are more users online that would be able to offer assistance to help any user.


58/72

50

6. References

1. Bell, Theo The Numerical Wind Tunnel: A Three-dimensional Computational

Fluid Dynamics Tool, M.S. Thesis, Dalhousie University, Halifax, Nova Scotia,August 2003.

2. Tinoco, Edward N. The Changing Role of Computational Fluid Dynamics in

Aircraft Development, AIAA Paper 98-2512, pp. 161-174, Boeing Commercial

Airplane Group, Seattle, Washington, 1998.

3. Fujii, Kozo and Miyaji, Koji WEB-CFD and Beyond CFD for non-CFD

Researchers, AIAA Paper 02-14233, 2002.

4. Pope, Alan Basic wing and airfoil theory, New York, McGraw-Hill, 1951.

5. OpenFlower CFD Software Reference Manual, OpenFlower Team, July, 2004.

6. NASA Glenn Research Center, http://www.sgevolution.com/htm/nasa.htm

7. NASA Advanced Supercomputing Division: Unsteady Flow Analysis Toolkit,

http://www.nas.nasa.gov/Software/UFAT8. JavaFoil Analysis of Airfoils, http://www.mh-aerotools.de/airfoils/javafoil.htm

9. NASA Glenn Research Center FoilSim, http://www.grc.nasa.gov/WWW/K-

12/airplane/foil2.html

10. KU Large Wind Tunnel Operating Handbook.

11. Deasi, S. S. Relative roles of computational fluid dynamics and wind tunnel

testing in the development of aircraft. Current Science, Vol. 84, No. 1, 10

January 2003.


59/72

51

12. W.H. Mason, D. L. Knill, A.A. Giunta, B. Grossman and L.T. Watson Getting

the Full Benefits of CFD in Conceptual Design AIAA Paper 98-2513, 16th

AIAA Applied Aerodynamics Conference June 15-18, 1998.

13. van Leer, Bram CFD Education: Past, Present, Future AIAA Paper 99-0910,

37th AIAA Aerospace Sciences Meeting and Exhibit, January 11-I 4, 1999.

14. Lamar, John E., Obara, Clifford J., Fisher, Bruce D., Fisher, David F. Flight,

Wind-Tunnel, and Computational Fluid Dynamics Comparison for Cranked

Arrow Wing (F-16XL-1) at Subsonic and Transonic Speeds NASA/TP-2001-

210629.15. FLUENT FlowLab 1.2 Documentation, Users Guide, Fluent, Inc, Jan 2005.

16. FLUENT 6.2 Documentation, Users Guide, Fluent, Inc, 2005.

17. FLUENT GAMBIT 2.2 Documentation, Users Guide, Fluent, Inc, 2005.


60/72

52

Appendix A: Program Flowchart of LabVIEW for LSVWT ProgramLabVIEW details of the programmingProgram Hierarchy


61/72

53

Main menu


62/72

54

Launch JavaFoil


63/72

55

Launch JavaFoil (continue)


64/72

56

Launch FoilSim


65/72

57

Launch FlowLab


66/72

58

2-D Panel Method


67/72

59

Airfoil pressure distribution chart


68/72

60

Airfoil coordinates conversion from generated to LabVIEW use


69/72

61

MATLAB source code for Panel Method used in LabVIEW

% Panel Code in MATLAB%% Open a File and read airfoil coordinates

%fid = fopen('panel.data.txt','r')%% Read Angle of Attack%alpha = fscanf(fid,'%f',1);%% read number of points on the upper side of airfoil%nu = fscanf(fid,'%d',1);%

% read number of points on the lower side of airfoil%nl = fscanf(fid, '%d',1);%% read Flag that states if this airfoil is symmetric% if isym > 0 then airfoil is assumed symmetric%isym = fscanf(fid,'%d',1);%% Read a scaling factor% The airfoil y- ordinates will be multiplied by this factor%factor=fscanf(fid,'%f',1);

if(isym>0)nl = nu;

end%% Allocate storage for x and y%x = zeros(1,100);y = zeros(1,100);%% Read the points on the upper surface%for i = nl:nl+nu-1

a=fscanf(fid,'%f',1); b = fscanf(fid,'%f',1);x(i) = a;y(i) = b * factor;


70/72

62

endif isym == 0%% If the airfoil is not symmetric, read lower side ordinates too..%

for i = 1:nla=fscanf(fid, '%f',1); b = fscanf(fid, '%f', 1);x(nl+1-i) = a;y(nl+1-i) = b * factor;

endelse

for i =1:nlx(nl+1-i) = x(nl-1+i);y(nl+1-i) = - y(nl-1+i);

endendfclose(fid);%% Plot the airfoil on window #1%% plot(x,y);n=nu+nl-2;A=zeros(n+1,n+1);ds=zeros(1,n); pi=4. * atan(1.0);%% Assemble the Influence Coefficient Matrix A%for i = 1:nt1= x(i+1)-x(i);t2 = y(i+1)-y(i);ds(i) = sqrt(t1*t1+t2*t2);

endfor j = 1:na(j,n+1) = 1.0;for i = 1:nif i == ja(i,i) = ds(i)/(2.*pi) *(log(0.5*ds(i)) - 1.0);

elsexm1 = 0.5 * (x(j)+x(j+1));ym1 = 0.5 * (y(j)+y(j+1));dx = (x(i+1)-x(i))/ds(i);dy = (y(i+1)-y(i))/ds(i);t1 = x(i) - xm1;


71/72

63

t2 = y(i) - ym1;t3 = x(i+1) - xm1;t7 = y(i+1) - ym1;t4 = t1 * dx + t2 * dy;t5 = t3 * dx + t7 * dy;

t6 = t2 * dx - t1 * dy;t1 = t5 * log(t5*t5+t6*t6) - t4 * log(t4*t4+t6*t6);t2 = atan2(t6,t4)-atan2(t6,t5);a(j,i) = (0.5 * t1-t5+t4+t6*t2)/(2.*pi);

endenda(n+1,1) = 1.0;a(n+1,n) = 1.0;end%% Assemble the Right hand Side of the Matrix system

%rhs=zeros(n+1,1);alpha = alpha * pi /180;xmid=zeros(n,1);for i = 1:nxmid(i,1) = 0.5 * (x(i) + x(i+1));ymid = 0.5 * (y(i) + y(i+1));rhs(i,1) = ymid * cos(alpha) - xmid(i) * sin(alpha);

endgamma = zeros(n+1,1);%% Solve the syetm of equations% In MATLAB this is easy!%gamma = a\rhs;cp=zeros(n,1);cp1=zeros(n,1);%% Open a file to write x vs. Cp and the Loads%% Change the file name below, to open a new file every time%fid=fopen('cp.data.txt','w');fprintf(fid,' X CP\n\n');for i = 1:ncp(i,1) = 1. - gamma(i) * gamma(i);cp1(i,1) = - cp(i,1);xa = xmid(i,1);cpa = cp(i,1);%


72/72

% Write x and Cp to the file%% The xa- coordinate is the center points of panel 'i'% Cpa is the Cp value at that point%

fprintf(fid,'%10.4f %10.4f\n',xa,cpa);end%% Open a new figure and plot x vs. Cp%%figure(2);%plot(xmid,cp1);%% Compute Lift and Drag Coefficients%cy = 0.0;

cx = 0.0;cm = 0.0;% We assume that the airfoil has unit chord% we assume that the leading edge is at i = nl;for i=1:ndx = x(i+1) - x(i);dy = y(i+1) - y(i);% xarm is the moment arem , equals distance from% the center of the panel to quarter-chord.xarm = 0.5 * (x(i+1)+x(i))-x(nl)-0.25;cy = cy - cp(i,1) * dx;cx = cx + cp(i,1) * dy;cm = cm - cp(i,1) * dx * xarm;end%% Print Lift and Drag coefficients on the screen%cl = cy * cos(alpha) - cx * sin(alpha)cd = cy * sin(alpha) + cx * cos(alpha)cmcpx=x'y=y'%

Documents

Low Speed Virtual Wind Tunnel Simulation